© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 24576
    9. 8-RES_FS24576-JB
    11. 6740
    Prev
    Next
    INDEX
    Explanations

    negative statements or criticism about political figures, especially related to Donald Trump

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs24576-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    Data Type
    torch.float32
    Hook Point
    blocks.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    314
    +0.16
    0.8%
    650
    +0.09
    0.4%
    530
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    314
    +0.16
    0.01
    530
    +0.09
    0.02
    477
    +0.09
    0.02
    Negative Logits
     GOODMAN
    -0.70
    anca
    -0.64
    ï¸ı
    -0.64
     strips
    -0.63
     bleach
    -0.59
     skim
    -0.58
     pens
    -0.58
     marrow
    -0.58
     Dickinson
    -0.57
     resemb
    -0.57
    POSITIVE LOGITS
    elect
    1.18
    design
    0.96
    appointed
    0.96
    driven
    0.88
    Elect
    0.86
    imposed
    0.84
    funded
    0.80
    govern
    0.79
    election
    0.78
    turned
    0.78
    Activations Density 0.041%

    No Known Activations